the inside game

What Democrats and Pollsters Missed in New Jersey

One of the state’s top experts gets candid.

The watch party for Republican Jack Ciattarelli, who nearly toppled New Jersey governor Phil Murphy. Photo: Spencer Platt/Getty Images
The watch party for Republican Jack Ciattarelli, who nearly toppled New Jersey governor Phil Murphy. Photo: Spencer Platt/Getty Images

So what just happened in New Jersey? After incumbent Democratic governor Phil Murphy scraped his way to reelection by less than two points on Tuesday, the question is top of mind not just in his state but for his party around the country. On one level, the answer is obvious: Democratic candidates paid the price for their party’s and Joe Biden’s unpopularity — it happened in Virginia, too, where former governor Terry McAuliffe lost to Republican Glenn Youngkin after leading almost the entire race.

But New Jersey’s near-reversal was more dramatic and much more surprising: For one thing, Biden just won it by 16 points last year. And for another, most polling showed Murphy beating Republican ex-state legislator Jack Ciattarelli by double digits until the final days, and even then, none of the tighter polls had the race even close to tied. Murphy seemed certain to outrun the national headwinds as a low-key progressive with high marks throughout the year for taking the state’s brutal COVID-19 outbreak seriously.

In July, Monmouth University poll director Patrick Murray told me “it would really take some type of implosion in terms of the governor’s job approval ratings” to make the race close. For a Republican to have a shot statewide, Murray said, the implosion “would have to be on an apocalyptic level.” I spoke with Murray again on Thursday to try and make sense of that apocalypse.

Let’s start with the top line. When you look at the overall numbers in New Jersey, what’s your conclusion on how Murphy almost lost? And given that the polling was wrong yet again, what was most surprising?

The most shocking thing was the huge increase in the Republican turnout in the southern part of the state. These are the working-class voters that Democrats have counted on in the old days, and they were sending a message that they felt that the party has just disappeared from them. Many of these are folks who supported Joe Biden just last year — they were never fans of Donald Trump — but they’re certainly not fans of what they see as a Democratic Party that’s obsessed with pleasing every progressive whim.

But the polling totally missed this? You — Monmouth — put out a range of possible results in your polling to account for different turnout possibilities, but the closest one still had Murphy up eight. Was this the biggest miss?

Yeah. It’s really interesting, because if you look at the polling in New Jersey and Virginia, Virginia polling was right on target, and New Jersey polling significantly underestimated Republican strength. [Under one model, Monmouth’s final Virginia poll showed an even race; Youngkin won by three points.] Four years ago, it was the opposite — or at least different. New Jersey polling was right on the money, and Virginia polling significantly underestimated Democratic strength. There are surges that are happening right now that are making it much more difficult to do election polling.

Let me back up and say election polling always has its shares of hits and misses, and good pollsters have more hits than misses. But we’re getting, frequently, many more misses than we should. Are we going to have any credibility in saying we’re accurately portraying what’s going on out there in the election? It undermines what we know is good work with all the public policy work we do. It’s one of the reasons why folks like Gallup and Pew got out of the election-polling business.

Presumably you’ll dig into this still in your audits, but do you have an answer for what accounts for the differences in accuracy between your New Jersey and Virginia polls?

No! No! In 2016, we were off, and we found a few things that really determined that. Some of it had to do with the sampling, some of it just simply had to do with the falloff in Hillary Clinton’s support at the very end, where unenthusiastic voters didn’t show up. We added that together and said, “Oh, we figured it out.” And sure enough: 2017, 2018, our polling was fine.

Then we hit 2020 and, again, it’s off. We went back and looked at our polling, went back and validated all the voters that we tried to contact, and we found that there was a significant chunk of voters who are likely to be Trump voters who would not talk to us for election polling. It’s not impacting our general population studies that we do with, say, COVID research and vaccination rates and all that stuff. It affects when we try to get down to a model of who an electorate is going to be.

And one of our working hypotheses was that maybe this was something that only had to do with Donald Trump. The results in New Jersey suggest, no, it’s not that — that there’s something going on with the culture wars right now that is carried over from Trump that is impacting our ability to make sure that we have accurate samples when we have these surge elections.

To be clear, you don’t see a conflict between the idea that this result is explained by this stealthily extra-activated Republican base and the notion that we simply saw traditional Republican voters who may have disliked Trump returning to the GOP fold for Ciattarelli?

What we saw is a number of voters who would not necessarily vote in a gubernatorial race coming out to do so because they were activated by national politics. It’s a kind of voter who really is activated by the culture wars, and that’s why we see this differential turnout in a particular area — which was friendly toward Donald Trump, but not particularly, overwhelmingly so. These are the kinds of working-class voters that Democrats have had for many years and have been losing now.

What we saw is a combination of two factors here: One was just a rejection of a focus on progressive politics. The other was die-hard Trump voters who would never vote for a Democrat anyway — but they normally would just not have voted in a gubernatorial election because they wouldn’t have cared enough — now starting to care. You put those two things together, and that’s the headwinds that Democrats are facing right now.

So what should we make of the commentary that this result is a backlash in particular to the school shutdowns under Murphy, and to his COVID handling more broadly? When we talked about this race in July, you made the point to me that his relatively high approval rating was a direct result of how he’d handled the pandemic, and how he’d taken it seriously —

Which was true in July, in August, and into early September. But when people said, “Oh, maybe this Delta thing isn’t as bad as we thought,” then that advantage evaporated.

By the way, the Murphy campaign was really bad at adjusting to this reality. They thought that if they could just hit away at all their progressive accomplishments over the last four years, that would be enough to swing the progressives to come out to vote. And here’s where we saw the other half of this problem. Turnout overall in the state was up by at least 15 points versus four years ago. Except in the core urban counties, where it’s running about even. These are the voters for whom Murphy’s message was, “I’ve given you all these things, aren’t I wonderful?” And they basically said, “No, not really, you’re not paying attention to what I really need.”

And you saw something similar in Virginia? Nationally?

Yes, right. It’s the pervading image that people have of the Democratic Party over the last year, where it’s heading. Murphy doubled down on that message, and so we have pretty clear evidence that that message is not resonating, and it’s not resonating among people they thought it was going to resonate with.

Earlier today, you tweeted, “Passaic. Passaic. Passaic?” That county swung aggressively toward the Republicans compared to 2017, though Murphy still narrowly won it. But can you explain to people who aren’t familiar with the state why you singled it out?

Passaic is a county that Democrats could generally solidly count on because of significant turnout. But it has an extremely high rate of immigrant population, and that raises some tensions there. So you have a combination of the tensions from people who will vote Republican because of those tensions, but you also have an immigrant population there that’s saying, “You keep telling me you’ve done all these wonderful things for me and that’s not what I want. I’m not concerned about immigration policy, I’m concerned about paying my bills, getting a job, and you’re not talking about that.” And so that’s where Passaic is very interesting. It has both groups in pretty equal measure. And both of them rejected that message, for different reasons.

Was the Ciattarelli campaign particularly savvy? Despite Murphy’s frequent attempts to tie him to Trump — like by repeatedly pointing out his attendance at a Stop the Steal rally — he eventually ran a pretty straightforward New Jersey Republican–style campaign, right?

Yeah, it was smart. Look, what they did was keep hammering away in their public messaging at the kind of issues that get Republicans elected statewide in New Jersey, which is taxes and affordability, and Phil Murphy being out of touch. That message wasn’t resonating for most of the campaign because COVID was just dominating. But once COVID dissipated as a priority, it opened up that window for people to start paying attention to that message and being reminded of it. “Oh, yeah, I don’t know who this Ciattarelli guy is, but Phil Murphy isn’t telling me he’s making the state more affordable for me, or that he plans to do that over the next four years.” The other part of it was he knew he needed those Trump voters to come out for him, he needed those culture-war voters to come out for him. But he didn’t need to do any messaging for that, because that message was handled nationally.

To take a step back for a second, I have been hearing some people — quietly — arguing that we shouldn’t read too much into the results in either state since they can largely be explained by the national atmosphere, plus there’s always a backlash to the party in power after a presidential election.

It’s wishful thinking. This was a big message, a huge message. We’re not just looking at Virginia and New Jersey, we’re looking at Long Island and Bucks County and a whole host of other areas.

The one thing that I will say is interesting in our findings in New Jersey — and I haven’t looked as closely in Virginia ones — is that those formerly moderate Republican suburban counties that swung Democratic during the Trump years held fairly steady. Those college-educated voters who rejected Trump and the direction the Republican party had taken are stuck there. They may not be overly thrilled with where the Democrats are going, but they are still unhappy with the direction Republicans are taking. Which is different from the remaining working-class Democrats and the urban voters, two very different groups. Working-class Democrats were drifting away from the Democrats anyway; the urban voters are the ones that political elites keep saying they’re helping. And the message was, “No, you’re not.” They need to go back and listen to their base; they don’t even have their base right.

With all the important caveats about how exit polls can be extremely unreliable, and acknowledging that there were some inconsistent findings on this question on Election Night, there is still some worry among Democrats that Latino voters are still shifting away from them, at least on the margins. Have you seen evidence of a real shift?

No, I think it’s one of the things where polling is not doing a very consistent job — with Latino voters. When you look at a place like Patterson, or votes in Hudson County, where there’s a huge Hispanic vote, there’s no question that the Democrats underperformed with those groups. The polls don’t seem to be capturing it as much, and I think the elite leadership who claim to represent the interests of those groups, who have the ear of the Democratic Party, probably are misrepresenting what these folks are concerned about. The majority are not concerned about immigration; they’re concerned about paying their bills.

Finally, there’s been a lot of understandable interest in some of the local results in New Jersey, especially in the race where state senate president Steve Sweeney lost to a truck driver no one had ever heard of before. Did the local races tell you anything different or surprising, or —

No. The local results were a product of the national mood. Sweeney is unlucky, in that he is a working-class Democrat who has been successful in the past saying there’s room for us and our type in the Democratic Party. And the message this year, that these voters were getting, was, “No, there’s not.”

This interview has been edited and condensed.

What Democrats and Pollsters Missed in New Jersey